Author’s Response

To the Editor:

Thank you for giving us the opportunity to respond to Dr. Udell’s thoughtful concerns regarding our recent article. As his letter indicates, this was indeed a retrospective study not of identifiable individual patients, but rather those persons who had claims submitted for diabetic lower-extremity ulcers during the years 2002 to 2003. Furthermore, his criticism is indeed correct regarding both the value of retrospective studies as well as their weaknesses in attributing causation or association with outcomes, which in this case pertains to amputations. In our type of retrospective (ecological) study of claims data, we do not have data on individual patients’ comorbidities, compliance, provider expertise, nor specifics on alternative therapies other than vacuum-assisted closure (VAC) therapy. This is indeed a weakness of these types of surveys; however, our strength lies in the numbers of patients under study in both the control and study groups. From such numbers, trends certainly do emerge and inferences can be made and corroborated with prospective study designs.

The point made about the end point of amputations and the variability in surgeons’ criteria for amputation is also valid. However, our end point is the only “hard” end point that could be meaningfully extracted from the datasets (other than death, not meaningful for the purposes of our study). While we agree that different surgeons will select different patients and different scenarios for subsequent amputation or time to amputation, it is also true that this very circumstance mimics the realities of practice and is not a controlled parameter. In fact, the literature is replete with numerous studies in which an amputation was a final end point of interest and all such studies also are subject to the vagaries of a surgeon’s determination. Another problem with a truncated dataset is the possibility of missing an amputation that might have occurred even 1 week after the final date of data collection. This, however, would be a random event that could be expected to occur in the same frequency as that found during the formal study period.

Our results support the contention that negative-pressure therapy treatment for diabetic foot ulcers significantly reduces the rate of amputation in the Medicare population compared to other traditional therapies. Although not a significant difference with the numbers studied, the same trend was evident in the commercial payer sample. Equally important, we found a significant decrease in the rate of amputations in the groups treated with negative-pressure wound therapy in both patient samples as depth of wound debridement proceeded from skin to bone. We also believe that these data mirror the experience found in clinical practice, and we welcome the publication of further prospective research on negative-pressure wound therapy.